Environmental science for the European refining industry Volume 24, Number 1 • February 2015
Environmental science for the European refining industry
Volume 24, Number 1 • February 2015
Considerable efforts have been made to assure the accuracy and reliability of the information contained in thispublication. However, neither Concawe nor any company participating in Concawe can accept liability for any loss,damage or injury whatsoever resulting from the use of this information. The articles do not necessarily represent theviews of any company participating in Concawe.
Reproduction permitted with due acknowledgement.
Cover photograph reproduced courtesy of ©iStockphoto.com
Editor: Robin Nelson, Concawe
Design and production: Words and Publications • [email protected]
C lean fuels have made a profound impact on
modern society, transporting people, both for
work and for leisure activities, and goods that have
enabled our living standards to improve with a con-
comitant increase in life expectancy. However, we
cannot rest on the achievements of the past and must
move on to new challenges. The petroleum refining
industry recognises it has a role to play, in reducing
the impact man has on the environment via reduced
GHG emissions, improving air quality in cities and using
precious natural resources in the most efficient way.
The articles in this Review report on ongoing work by
Concawe to further contribute to the safety of everyone
working in our industry, to improve the environmental
performance in the manufacture of fuels and other oil
products and to develop the fuels for the future, nec-
essary to maintain the position of oil-derived liquid
fuels as the most efficient option until fully renewable
transport fuels become widely available at an afford-
able price.
Foreword
Robin Nelson
Science Director
Concawe
1Volume 24, Number 1 • February 2015
Contents
Concawe review2
Abating fugitive VOC emissions more efficientlyComparing best available techniques for detecting refinery fugitive emissions 3
This article summarizes the Concawe study, ‘Techniques for detecting and quantifying fugitive emissions—results of
comparative field studies’, that compares the two BAT (best available techniques) detection methods for refinery
fugitive emissions (leaks) of non-methane volatile organic compounds (NMVOC): ‘sniffing’ and ‘optical gas imaging’
(OGI). The main finding is that the OGI technology is faster and can effectively detect the main leaks. By repairing
those leaks, a reduction comparable to that achieved using the sniffing method is achieved, contributing to the total
site NMVOC emission reduction.
Enquiries to: [email protected]
Spark versus compression ignition in a new energy environmentCould new engine technologies help to address the diesel/gasoline supply imbalance? 10
Dieselisation of the fuel market is accelerating as commercial transport increases and as the fuel consumption of
new passenger cars decreases. At the same time, the recent US revolution in shale gas and tight oil is putting new
pressure on worldwide refining as more gasoline molecules are available from light crude extraction and from
chemical feedstock substitution. From a renewable fuel perspective, ethanol and its derivatives are available in
larger volumes.
As the world considers how to achieve best-in-class ‘well to wheels’ greenhouse gas reduction from transport,
is it time to look at new opportunities to combine the highest efficiency engine technology with the lowest
greenhouse gas (GHG)-emitting fuels? Would a low octane, low cetane gasoline or even a current pump gasoline
be a better choice for future compression ignition engines?
Enquiries to: [email protected]
Downstream oil industry safety statistics for 2013Concawe’s latest analysis of personal injury and process safety data 17
Safety management systems are widely recognized by the oil industry as an essential tool for collecting and analysing
safety incident data and continuously improving the safety of employees and contractors. To support this effort,
Concawe has, since 1993, been compiling statistical safety data for the European downstream oil industry in order to:
1. provide member companies with a benchmark against which to compare their own company’s safety performance; and
2. demonstrate how responsible approaches to safety management can help to ensure that accidents stay at low
levels in spite of the hazards that are intrinsic to refinery and distribution operations.
Most importantly, Concawe’s annual safety data report enables companies to evaluate the efficacy of their own
management systems, identify any shortcomings, and take corrective actions as quickly as possible.
Enquiries to: [email protected]
Abbreviations and terms 21
Concawe contacts 22
Concawe publicationsReports published and co-authored by Concawe 23
This article summarizes the Concawe study,
‘Techniques for detecting and quantifying fugitive
emissions—results of comparative field studies’, that
compares the two BAT (best available techniques)
detection methods1 for refinery fugitive emissions
(leaks) of non-methane volatile organic compounds
(NMVOC): ‘sniffing’ and ‘optical gas imaging’ (OGI).
The main finding is that the OGI technology is faster
and can effectively detect the main leaks. By repairing
those leaks, a reduction comparable to that achieved
using the sniffing method is achieved, contributing to
the total site NMVOC emission reduction.
The petroleum refinery industry has successfully reduced
NMVOC emissions—one of the precursors to surface
level ozone formation—through leak detection and repair
(LDAR) programmes, and technology advances (e.g.
improved valve packing). To go further with this reduc-
tion, the industry is now focusing its efforts on the control
of fugitive emissions which can contribute up to one third
of the total site NMVOC emissions. Fugitive emissions
are generated at plant components which are supposed
to be leak-tight (e.g. pump or compressor seals, valve
packings, flanges, sample points, etc.). While a typical
site would have more than 50,000 such components,
only a few of these contribute to the bulk of fugitive emis-
sions. Identifying these leaks for repair is difficult and time
consuming, as they will be spread out over the entire
site, and in locations which are difficult to access.
Two methodologies are currently available to detect
leaking equipment in LDAR programmes:
1. ‘Method 21’ (or ‘sniffing’), developed by the
US-EPA, involves the use of a hydrocarbon ionisa-
tion detector; it was historically the first approach
and is a widely accepted method.
2. Optical gas imaging, using an infra-red camera, is a
newer technique which is gaining increasing
acceptance.
Both methods are effective, and each has advantages
and limitations (outlined below). However, as they are
based on different technologies and applied in the field
in different ways, a comparison is not straightforward.
The two methods had not previously been compared in
large simultaneous, independent field trials in Europe.
Such trials were the objective of a project managed by
the Concawe OGI Group, and the results obtained are
summarized in this article.
Background
Initial methodology: ‘Method 21’ or ‘sniffing’
The monitoring and emissions estimating methodology
is described in EPA-453/R95-017 (US) and in
EN 15446:2008 (EU) (CEN, 2008), and is commonly
referred to as ‘sniffing’.
A hand-held hydrocarbon detector (either a flame ion-
isation detector (FID) or photo ionisation detector (PID))
is used to ‘screen’ all potential leak points one by one
and record, for each of them, the highest hydrocarbon
concentration measured (screening value). Above a
given concentration threshold (e.g. 10,000 ppmv), the
equipment is identified as leaking and must be
repaired. A maximum of around 500 components
(most of which will not be leaking) can be screened
effectively per work day by one person. In EN 15446,
factors are provided per equipment type and service,
to permit the estimation of NMVOC mass emissions
based on the screening value. Those factors were
derived by the EPA from a statistical analysis of a sig-
nificant number of leaks from various components,
which were simultaneously both screened and
‘bagged’ (i.e. the leak flow is captured in an imperme-
able bag and its concentration and composition are
analysed, allowing its emitted mass to be calculated).
These data showed a very large spread, e.g. for the
same hydrocarbon concentration, the mass emission
could vary by as much as four orders of magnitude.
For the lower concentrations, correlations were devel-
oped on a log/log scale but can only be statistically
representative if applied to a very large number of
components. For the higher concentrations (above
50,000 ppmv methane) the FID and PID detectors do
3Volume 24, Number 1 • February 2015
Comparing best
available techniques
for detecting refinery
fugitive emissions
Abating fugitive VOC emissions more efficiently
1 The ‘sniffing’ method and the optical gas imaging technique are part of BAT6: monitoring of diffuse VOC emissions to air.
not give a linear response. Therefore, the methodology
assigns ‘pegged values’—fixed mass emission values—
to the high concentration readings (e.g. >10,000 ppmv
and > 100,000 ppmv).
Newer technology: optical gas imaging (OGI)
In OGI technology, passive mid-wave infrared cameras
are equipped with a filter to selectively detect radiation
at the specific C-H absorption band (3.2–3.4 µm). The
commercial OGI cameras are easy to use and show the
hydrocarbon leak as a plume coming from the emitting
source. OGI can detect any leak whereas sniffing can-
not survey components which are not accessible. A
major advantage of OGI is the monitoring speed. The
OGI technology provides a qualitative assessment of
the size of the leak. The main limitation of OGI is its
higher minimum detection limit, i.e. 1–10 g/h, depend-
ing on the hydrocarbon, compared to about 0.01 g/h
for sniffing.
OGI has proven to be very useful in safety and mainte-
nance applications, and is now commonly used after a
unit start-up to verify equipment tightness.
The latest camera models on the market are the FLIR
GF320 and the OPGAL EyeCGas. Based on the feed-
back from several contractors performing OGI surveys,
they give comparable results in the field.
Effectiveness of LDAR programmes
Over the years, LDAR programmes based on sniffing
have helped to reduce fugitive emissions. Data now
available indicate that OGI-based programmes would
most likely have achieved similar reductions faster and
cheaper, considering that only the largest leakers (less
than 2% of the total equipment population) are respon-
sible for more than 90% of the fugitive mass emissions
(API, 1997).
Europe promoting the use of OGI-based LDAR
The latest BAT Reference document (BREF) for refining
of mineral oil and gas (REF BREF) considers both sniff-
ing and OGI as BAT. In 2013, the Netherlands
Abating fugitive VOC emissions more efficiently
Standardization Institute (NEN) developed guidelines for
performing OGI surveys, aimed at providing a common
methodology (National Technical Agreement 8399:2013)
(NEN, 2013).
Concawe study objective
In 2012–2013, Concawe carried out several parallel
LDAR campaigns. Both OGI and sniffing (EN
15446:2008) were applied by two independent teams.
The objective was to compare the NMVOC mass emis-
sions detected by each method. The mass emissions
were independently estimated for all detected leaks by
‘bagging’2, when possible. The bagging technique
applied uses a combination of two instruments: the
High Flow® Sampler (a device developed by manufac-
turer Bacharach for estimating natural gas leaks) and
the ‘TVA-B’—a FID/PID detector commonly used in
sniffing surveys. The High Flow® Sampler was used to
estimate the volumetric flow rate of the leak. The TVA-B
was used at the outlet of the High Flow® Sampler to
estimate the VOC concentration of the leak. The com-
bination of these two techniques, which is much faster
than the original methodologies described in
EPA-453/R95-017, will be referred to as ‘HFS’ through-
out this article.
HFS was validated in a controlled experiment and com-
pared to the EPA ‘vacuum bagging’ method in the field
(20 leaks were bagged by both methods). The limit of
the validation resulting from the controlled experiment
was 200 g/h and this was used as a maximum HFS rate
when analysing the results of the field campaign. This
approach is similar to the ‘pegged values’ in Method 21
(see above). For the leak rates between 20–200 g/h,
HFS was found to give a larger leak rate by a factor 2 to
5 than vacuum bagging. However, as this results in a
conservative estimation of the NMVOC emissions, the
HFS results were used for leaks between 20–200 g/h in
this study. For the lowest leak rates (1–20 g/h) HFS
accuracy was comparable to vacuum bagging.
Concawe review4
2 Bagging techniques are not applicable for regular LDAR surveys as only a maximum of 20 leaks per day can be bagged.
Parallel sniffing and OGI surveys
Units handling gas and light hydrocarbons were sur-
veyed by both methods at two European refineries.
Site 1 is a newer facility (built in the 1980s) where LDAR
was applied for the first time during this survey. Site 2 is
an older facility with an LDAR programme in place for
10 years. A single campaign was done at Site 1
(November 2012) while three consecutive campaigns
were done at Site 2 (between June and November
2013). In the first campaign at Site 2 several units were
surveyed, totalling 25,000 LDAR points. In the subse-
quent campaigns, only sub-unit 1 was surveyed
(selected as previous surveys had shown this to have a
relatively high number of leakers). Site 1 and Site 2 sub-
unit 1 have approximately 4500 LDAR points each. The
leak definition threshold was 10,000 ppm for Site 1 and
5000 ppm for Site 2 (based on the site permit).
Experience with sniffing has shown that the number of
components classified as ‘leakers’ does not increase
significantly when the leak threshold definition drops
from 10,000 ppm to 5000 ppm, and the two sites can
still be compared.
To improve the comparison for site 1 the bagged leaks
that were below the Site 1 leak definition but were
above or close to the Site 2 leak definition are added to
the analysis. In this Concawe work, a leak is defined as
either a visible OGI image or a screening value above
site leak definition.
The OGI surveys were performed according to the
Dutch guideline (NEN, 2013). The FLIR GF320 camera
was used and the equipment was surveyed at no more
than two metres distance from multiple angles (for the
accessible components). The pace of the survey was
2000 components per person per work day. The sniffing
surveys were performed according to EN 15446:2008.
The analyses for comparing the VOC mass emissions
estimated by the various methodologies were only
done for the bagged accessible leaks. This approach
was selected to make the comparison meaningful.
Method 21 correlations are only statistically meaningful
if applied to a very large number of leaks. The accuracy
of the Method 21 estimations for the number of leaks
detected in these partial surveys, therefore, is not as
high as when full site surveys are undertaken.
The main four observations made during the field LDAR
surveys are illustrated and discussed below.
1. The emissions estimated by the EN 15446 factors
and correlations are conservative for a facility where
no leaks above 200 g/h are present.
Two similar process units were surveyed in two different
European refineries (Site 1 and Site 2 sub-unit 1).
Figure 1 shows the number of leaks detected by sniffing,
and how many of those leaks had a screening value
Abating fugitive VOC emissions more efficiently
5Volume 24, Number 1 • February 2015
Figure 1 Total number of leaks found by sniffing (Site 1 and Site 2 sub-unit 1, campaign 3)
num
ber
of le
aks
140
100
80
60
40
0
20
Site 1 Site 2
120
non-pegged > 100,000 ppm
48
39
53
69
leak
rat
e (k
g/h)
6
5
4
3
2
0
1
Site 1 Site 2
HFS Method 21
0.3
3.7
3.3
5.4Remark: not all detected accessible leakscould be bagged (e.g. hot surface equipment)
Figure 2 Leak rates estimated by two methods (Site 1: 74 leaks; Site 2: 97 leaks)
above 100,000 ppm (pegged leak) and how many were
below 100,000 (non-pegged leak).
In the two facilities, the fraction of ‘pegged leaks’ was
comparable (45% in Site 1 and 57% in Site 2 sub-unit 1).
Figure 2 shows the mass of these leaks estimated with
Method 21 and with HFS.
While the number of ‘pegged leaks’ is comparable, as
shown in Figure 1, Site 1 has fewer leaks in total and no
single large leak (≥ 200 g/h) based on the bagging
results (HFS). Site 2 sub-unit 1 has more leaks in total
(but a lower leak threshold) and 8 large leaks.
For Site 2 sub-unit 1, the emissions estimated with
Method 21 are close to those estimated with HFS (a
factor of 1.6 difference). For Site 1, the emissions esti-
mated with Method 21 are much higher than the HFS
estimation (a factor of 12 difference). A possible expla-
nation is that the Method 21 factors and correlations
were established many years ago, when the occur-
rence of large leaks was statistically more frequent. This
method has not been revised in 20 years and could
misrepresent the current situation, where LDAR pro-
grammes and technology advances (e.g. improved
valve packing) have resulted in reduced fugitive emis-
sions relative to 20 years ago.
Abating fugitive VOC emissions more efficiently
2. OGI and sniffing may not find the exact same
leaks. However, the ‘common leaks’ found represent
the largest portion of the total VOC mass emissions.
Figures 3 and 4 show, for Site 1 and Site 2 sub-unit 1,
the number of leaks detected by the two methods and
the mass of these leaks (calculated with the HFS
method). As illustrated above, Site 1 and Site 2 sub-
unit 1 are very different in terms of total NMVOC mass
leak rate.
In Site 1 (Figure 3), the number of leaks only identified
by sniffing was significant (70 out of 104), but the mass
of these leaks (0.15 kg/h) is smaller than the mass of
the common leaks (0.18 kg/h). One could argue that
OGI ‘missed’ 0.15 kg/h of NMVOC mass on accessible
components, but the three ‘OGI-only’ leaks which
could not be quantified (non-accessible) are likely to
generate an equivalent mass emission to the ‘Method
21-only’ leaks.
In Site 2 sub-unit 1 (Figure 4), both the number and the
mass of common leaks are the most important. The
mass of ‘OGI-only’ leaks is comparable to the mass of
‘Method 21-only’ leaks.
Concawe review6
num
ber
of le
aks
80
70
60
40
20
0
10
Both identified Method 21 only
number of leaks detected number of leaks bagged
OGI only
mass of bagged leaks (kg/h)
50
30
0.20
0.14
0.12
0.08
0.04
0
0.02
0.10
0.06
0.16
0.18
mas
s of
leak
s (k
g/h)
26 213
70
52
0.18
0.15
Remark: not all leaks couldbe bagged (e.g. hot surfaceequipment or non-accessibleOGI leaks)
Figure 3 Site 1, leaks identified by detection method and mass ofbagged leaks (estimated with HFS)
num
ber
of le
aks
90
70
60
40
20
0
10
Both identified Method 21 only
number of leaks detected number of leaks bagged
OGI only
mass of bagged leaks (kg/h)
50
30
3.5
3.0
2.0
1.0
0
0.5
2.5
1.5
mas
s of
leak
s (k
g/h)
8371
3926
3.1
0.3
80
18 17
0.7
Remark: not all leaks could bebagged (e.g. hot surface equipmentor non-accessible OGI leaks)
Figure 4 Site 2, sub-unit 1: leaks identified by detection method andmass of bagged leaks (estimated with HFS)
3. OGI was able to detect up to 90% of the total
NMVOC mass of accessible leaks in a single cam-
paign. This is comparable to sniffing, where some
leaks are missed (e.g. where equipment is not
accessible or is missing from the LDAR database).
Figure 3 shows that, for Site 1, the mass of OGI leaks
quantified is 55% of the total mass of accessible leaks.
Figure 4 shows that, for Site 2 sub-unit 1, the mass of
OGI leaks is 90% of the total mass, which is in line with
an analysis done in 1997 (Lev-On et al., 2007) by the
American Petroleum Institute (API). OGI effectiveness is
highest when the fugitive emissions from a facility are
relatively high: total NMVOC mass emission in Site 2
sub-unit 1 is 11 times higher than in Site 1 (3.3 kg/h
versus 0.3 kg/h for a comparable process and size, as
shown in Figure 2 on page 5). When the facility has rel-
atively low fugitive emissions, e.g. Site 1, the effective-
ness of OGI is lower but comparable to Method 21.
Figure 5 shows, for Site 2 sub-unit 1, three successive
OGI campaigns performed over six months. A very small
number of leaks were repaired between the campaigns
(only those with a potential safety issue). Successive
campaigns show that some additional leaks were found
and some previous leaks were not detected again. An
unexpected shut-down took place between campaigns
2 and 3; the opening of some equipment could explain
the higher number of new leaks in campaign 3.
In the same way, successive sniffing campaigns also
point out differences in the leak screening values. But
as OGI surveys are faster, it is possible to increase cam-
paign frequency at similar cost and improve leak detec-
tion effectiveness.
4. In real conditions, the OGI detection limit cannot
be defined by one single number. For the Concawe
survey (Site 2 sub-unit 1, Campaign 3), OGI
detected all leaks above 43 g/h and 80% of the leaks
above 1 g/h (out of all leaks bagged with HFS).
Figure 6 on page 8 shows all the third campaign bagged
leaks in the Concawe survey on a log/log scale. The
x-axis is the sniffing concentration while the y-axis is the
NMVOC mass flow, estimated using HFS. Two horizon-
tal lines can be drawn dividing the data into three zones:
all the leaks in the top section were detected by OGI
(> 43 g/h); most of the leaks in the middle section were
also detected by OGI (between 1 and 43 g/h); leaks in
the bottom section (below 1 g/h) were difficult, but not
impossible under ideal conditions, to detect with OGI.
In the middle section of Figure 6 (referred to as the ‘par-
tial OGI leak detection zone’), there were 90 leaks
bagged, with an average emission rate of 13.6 g/h.
Twenty-four leaks were missed by OGI and 11 leaks
were missed by sniffing.
Estimation of NMVOC mass emissions whenusing OGI
For OGI, the plume image only gives qualitative informa-
tion of the leak size. In 2004, the API published leak/no-
leak factors to be used in OGI campaigns to report
NMVOC mass emissions (see Table 1 on page 8). These
factors are based on a model refinery with a statistically
relevant leak population, surveyed by OGI. For modelling
the leak behaviour, the same bagging data were used as
in Method 21. The factors were developed for four differ-
ent lower detection limits of OGI cameras in the field.
Based on the observed ‘average’ field detection limit
for the new camera model FLIR GF320, when applied
according to the Dutch protocol (regarding distance
and survey speed), the leak/no-leak factors for 6 g/h
(leak definition, Table 1) were chosen for use in the
analysis of the field measurement data.
Abating fugitive VOC emissions more efficiently
7Volume 24, Number 1 • February 2015
leaks detected during 1st (and subsequent) campaign(s)
leaks detected during 2nd (and subsequent) campaign(s)
leaks detected during 3rd campaign
num
ber
of le
aks
dete
cted
by
OG
I
100
80
60
40
0
20
1st campaign 3rd campaign
120
89
2nd campaign
75
56
29
18
27
Figure 5 Site 2, sub-unit 1: leak trend by campaignfor OGI leaks
Figure 7 shows, for Site 1 and Site 2 sub-unit 1, a
comparison of the NMVOC mass emission (from
bagged leaks only) based on the different methodolo-
gies: Method 21, HFS and leak/no-leak factors (6 g/h
detection limit). The leak/no-leak factors give an over-
estimate of the emissions for Site 1, as does
Abating fugitive VOC emissions more efficiently
Method 21. They give a reasonable estimate for Site 2.
Knowing that the fugitive NMVOC emissions for Site 1
and Site 2 sub-unit 1 are very different, illustrating the
variability that can occur between facilities, the choice
of the API leak/no-leak factors for a 6 g/h leak defini-
tion seems reasonable.
Concawe review8
Table 1 Leak/no-leak factors for OGI surveys (API, 2004)
Figure 6 Site 2, sub-unit 1: OGI detection sensitivity
mas
s em
issi
on b
y H
FS (g
/h)
100
10
1
screening value (ppmv)
5000 50,000 No OGI leakdetection zone
Partial OGI leakdetection zone
OGI leakdetection zone
OGI and sniffing sniffing only OGI only
OGI hard detection = 43.2 g/h
OGI partial detection average emission rate = 13.6 g/h
OGI lower detection limit = 1.2 g/h
Component type Emission factors (g/h) for specified leak definitions
Leak definition—instrument detection 3 6 30 60limit (g/h)
Valves No leak 0.019 0.043 0.17 0.27
Leak 55 73 140 200
Pumps No leak 0.096 0.13 0.59 0.75
Leak 140 160 310 350
Flanges No leak 0.0026 0.0041 0.01 0.014
Leak 29 45 88 120
All components No leak 0.007 0.014 0.051 0.081
Leak 56 75 150 210
Conclusion
The Concawe parallel surveys, based on four large field
trials, confirmed that sniffing and OGI are equally able
to detect fugitive NMVOC emissions. OGI provides a
better identification of the leaks with a high mass emis-
sion. The OGI detection limit has improved in the past
few years: the new camera models are now able to
detect leaks of a few g/h with a high probability. The
leaks from accessible components not detected by
OGI are all small in size and represent a small fraction of
the total NMVOC mass emissions. OGI has the advan-
tage over sniffing of being able to detect any leak above
the detection limit present in the surveyed area, and not
only the leaks from accessible components listed in the
site database. OGI surveys also have the advantage of
being much faster, allowing more frequent surveys than
sniffing at comparable cost. For the OGI surveys using
the new camera models at the surveyed refinery sites,
the API leak/no-leak factors for a 6 g/h leak definition
provided a reasonable, although conservatively high,
estimate of the VOC mass emissions.
In a forthcoming Concawe report detailing this study, an
LDAR survey protocol will be proposed using OGI as a
standalone technique, comprising both detection and
quantification (estimation). This protocol will have a
detection efficiency of fugitive emissions similar to the
sniffing programmes currently practiced in Europe.
Looking ahead
An attempt to improve the existing OGI quantification
factors based on new leak bagging and statistical
analysis is not justified because the assumptions
needed to derive statistical correlations will at best rep-
resent an ‘average’ site situation. The methods for esti-
mating actual NMVOC emissions, e.g. by bagging, are
time consuming and/or subject to inaccuracies.
Moreover, one should bear in mind that the main objec-
tive of LDAR surveys is to reduce fugitive emissions (by
identifying leaking components for repair). Only a tech-
nology step-out, e.g. an improved OGI camera allowing
direct and fast leak mass quantification, has the poten-
tial to substantially improve the estimation of fugitive
NMVOC emissions in the future.
References
CEN (2008). EN 15446:2008. Fugitive and diffuse emissions ofcommon concern to industry sectors. Measurement of fugitiveemission of vapours generating from equipment and pipingleaks. ICS 13.040.40. Comité Européen de Normalisation(European Committee for Standardization), July 2008.
API (1997). Analysis of Refinery Screening Data. API PublicationNo. 310; Washington, D.C.
NEN (2013). NTA 8399:2013 en. Air quality - Guidelines fordetection of diffuse VOC emissions with optical gas imaging. ICS13.040.20. Netherlands Standardization Institute, October 2013.
Lev-On, M., Epperson, D., Siegell, J. and Ritter, K. (2007).Derivation of new emission factors for quantification of massemissions when using optical gas imaging for detecting leaks. In:Journal of the Air & Waste Management Association (JAWMA),Volume 57, Issue 9, pp.1061–1070, September 2007.
Abating fugitive VOC emissions more efficiently
9Volume 24, Number 1 • February 2015
Figure 7 Comparison of the VOC mass emission basedon the different methodologies
leak
rat
e (k
g/h)
6
5
4
3
2
0
1
Site 1
HFS Method 21
0.2
1.3
leak/no leak 6 g/h
1.2
Site 2
3.1
4.75.1
Asia Pacific
30
25
20
15
10
5
02000 2020 2040
gasoline ethanol diesel biodiesel jet fuel fuel oil natural gas other
Europe
30
25
20
15
10
5
02000 2020 2040
North America
30
25
20
15
10
5
02000 2020 2040
fuel
mix
(MB
DO
E)
Dieselisation of the fuel market is accelerating as
commercial transport increases and as the fuel
consumption of new passenger cars decreases. At the
same time, the recent US revolution in shale gas and
tight oil is putting new pressure on worldwide refining
as more gasoline molecules are available from light
crude extraction and from chemical feedstock substitu-
tion. From a renewable fuel perspective, ethanol and its
derivatives are available in larger volumes.
As the world considers how to achieve best-in-class
‘well to wheels’ greenhouse gas reduction from trans-
port, is it time to look at new opportunities to combine
the highest efficiency engine technology with the lowest
greenhouse gas (GHG)-emitting fuels? Would a low
octane, low cetane gasoline or even a current pump
gasoline be a better choice for future compression igni-
tion engines?
Introduction
The overall world energy demand is evolving; demand
is increasing in developing parts of the world, while in
others, such as Europe and North America, it is declin-
ing. Looking more closely at transportation, as this is the
sector that accounts for about 60% of oil demand today,
Concawe review10
it is clear that an understanding of the trend in overall fuel
demand and the mix of fuel types over time is critical to
ensuring that refinery operations and trade opportunities
are effective in meeting the changing requirements of
people around the world efficiently over time.
It is predicted that, in the period to 2040, the growth in
demand for transportation will be led by the Asia
Pacific region, and demand in North America and
Europe will remain relatively flat while energy demand
in Europe for light-duty vehicles is expected to decline
by about 40%. Commercial transportation is likely to
grow by about 20%, keeping energy demand for trans-
portation stable.
Global diesel demand will increase by more than 75%,
led by the Asia Pacific region, where demand for diesel
will more than double. The demand for diesel fuel in
North America will increase by about 60% even as the
overall demand for transportation fuel remains relatively
flat. In Europe, diesel volumes grow (+14% by 2025)
and then decline (-11% from 2025–2040), remaining
stable over the outlook period (+2% until 2040).
Global demand for gasoline will be relatively flat from
2010 to 2040, led by a decline in North America. Gasoline
Could the development
of new, more efficient
engine technologies
also help to address
the diesel/gasoline
supply imbalance?
Spark versus compression ignition ina new energy environment
Figure 1 Transportation fuel mix by regionS
ourc
e: a
dapt
ed fr
om T
he O
utlo
ok fo
r En
ergy
:A
Vie
w to
2040
(Exx
on M
obil,
201
4).
demand declines due to improved light-duty fuel effi-
ciency as well as increased use of oxygenates. In Europe,
the demand for motor gasoline drops significantly (-32%
by 2040). It is predicted that oil will remain the fuel of
choice for transport in the coming decades, making up
89% of the transport fuel mix for Europe by 2040.
The demand for natural gas used for transportation is
expected to grow by nearly 70%, with 60% of the
growth in the Asia Pacific region and 15% in North
America. Most of this growth comes from commercial
transportation, such as liquefied natural gas (LNG) for
long-haul trucks and compressed natural gas (CNG)
fuelling local delivery fleets and buses.
The past few years has seen a boom in the production
of natural gas in the USA, due to both continued pro-
duction of conventional gas resources, as well as a
surge in unconventional gas. In Pennsylvania, for exam-
ple, the Department of Environmental Protection is
reporting a 15x increase in natural gas production. In
addition to this there has been an increase in tight oil
production in some states, for example North Dakota.
These trends are expected to continue.
Challenge to EU competitiveness
As a result, energy prices for US refineries have the
advantage relative to those in Western Europe, as well
as in the Asia Pacific region. Energy prices in Europe
are currently much higher than in the USA, making it
difficult for European businesses to compete on the
global stage.
EU prices are twice as high for electricity, and three
times as high for gas. This regional difference is
expected to remain large through to 2035. Energy costs
make up around 60% of European refineries’ total oper-
ating costs, versus 28–30% for Eastern US refineries.
In addition, crude production growth worldwide is
focused in the USA, Iran and Canada, with more mod-
est growth in the Middle East and Western Africa, and
essentially constant production in Europe. Taken
together, Western European refiners are disadvantaged
versus other global refiners in terms of energy prices,
and crude availability.
The increasing global diesel demand, the decreasing
demand for gasoline and increasing availability of gaso-
line-type molecules from light crude extraction and
chemical feedstock substitution, create a diesel-gaso-
line imbalance which is expected to increase as time
goes on. These factors, as well as pressures to reduce
GHG emissions, mean that refiners worldwide, and par-
ticularly those in Western Europe, are coming under
increasing pressure.
Focus on vehicle efficiency
On the other hand, as pollutant emissions from motor
vehicles continue to fall to meet lower regulated emis-
sion limits, attention is increasingly focused on vehicle
efficiency and on fuel consumption to address future
concerns with energy supplies and transport’s contri-
bution to GHG emissions. Engine, aftertreatment and
vehicle technologies are evolving rapidly to respond to
these challenges.
Considerable research is concentrating on improving
the combustion performance of light-duty engines.
Compared to spark ignition (SI) engines, compression
ignition (CI) engines are already very efficient so the
challenge is to maintain or improve CI engine efficiency
while further reducing pollutant emissions. Engines
using advanced combustion technologies are being
developed that combine improved efficiency with lower
Spark versus compression ignition in a new energy environment
The diesel-gasoline
imbalance is expected
to increase into the
foreseeable future.
11Volume 24, Number 1 • February 2015
©S
hutt
erst
ock.
com
engine-out emissions, thus reducing the demand on
exhaust aftertreatment systems and, potentially, vehicle
costs. Because these concepts combine features of
both SI and CI combustion, the optimum fuel charac-
teristics could be quite different from those needed by
today’s conventional gasoline and diesel engines.
In general, these advanced combustion concepts sub-
stantially homogenize the fuel-air mixture before com-
busting the fuel under low-temperature combustion
(LTC) conditions without spark initiation. These
approaches help to simultaneously reduce soot and
NOx formation. Light-duty diesel engines are well suited
for advanced combustion because the higher fuel injec-
tion pressures, exhaust gas recirculation (EGR) rates,
and boost pressures that aid conventional CI combus-
tion also enable future variations of advanced combus-
tion. In addition, the duty cycle of light-duty diesel
engines emphasizes lighter loads where advanced
combustion is most easily achieved. Many of the nec-
essary hardware enhancements exist today, although
they may be expensive to implement in production
engines. Nonetheless, advanced combustion engines
are rapidly moving from research into engine develop-
ment and commercialisation.
The gasoline compression ignitionconcept
From a commercial perspective, it is well understood
that there are significant challenges associated with
bringing both a new engine concept and a dedicated fuel
into the market at the same time. The potential benefits
of fuelling advanced CI engines with market gasoline
merits further consideration for the following reasons. In
general, CI engines have an efficiency advantage over SI
engines, and extending their capability to use a broader
range of fuels could be advantageous. Second, the abil-
ity of CI engine concepts to use an already available mar-
ket gasoline would allow these concepts to enter the
fleet without fuel constraints. Third, more gasoline con-
sumption in passenger cars would help to rebalance
Europe’s gasoline/diesel fuel demand on refineries and
reduce GHG emissions from the fuel supply. Fourth, a
successful GCI (‘gasoline compression ignition’) vehicle
could potentially compete in predominantly gasoline
markets in other parts of the world.
Spark versus compression ignition in a new energy environment
Because of these potential benefits, it was decided to
investigate more completely the GCI engine concept,
specifically to determine the range of conditions over
which an engine could operate successfully in CI mode
on a European market gasoline. In addition to an engi-
neering paper study and a bench engine study on the
GCI concept (Rose et al., 2013), computational fluid
dynamics (CFD) in-flow and combustion simulations
were also carried out (Cracknell et al., 2014).
Computational fluid dynamics is a state of the art sim-
ulation tool for analysing the flow behaviour in internal
combustion engines. To reduce the computation time,
only the combustion chamber and port geometries
were modelled in this study.
Two main areas which were modelled were gas
exchange and turbulent non-reacting flow which was
Concawe review12
An early single-cylinder gasoline compression ignition
(GCI) test engine developed by Delphi; ongoing
development of the GCI concept aims to achieve diesel-
like efficiency with low CO2 emissions under real-world
operating conditions.
Del
phi
modelled using STAR-CD®, a CFD programme concen-
trating on the combustion chamber and the port
geometries in conjunction with a 1-D GT-Power simu-
lation which was used to define boundary conditions.
The other area was combustion modelling which was
done using KIVA software. The KIVA package included
other sub-models for looking at spray, turbulence, wall
impingement and other aspects. The STAR-CD® fluid
flow was mapped to the KIVA simulations and used as
boundary conditions (pressure, temperature, gas com-
position and flow velocity (see Figure 2).
Engineering paper study
An engineering paper study was first completed to
analyse critical engine and fuel parameters and judge
what speed/load range might be feasible for a GCI
engine concept. For this engineering study, and for the
bench engine work that followed, it was assumed that
the GCI engine concept would be fuelled with a typical
European market gasoline. The only change that was
made was the addition of a lubricity additive.
The paper study identified the autoignition resistance of
market gasoline as the single most critical challenge,
particularly at low load conditions. Three main
approaches were identified to mitigate this challenge:
l Shortening the ignition delay by increasing the
charge pressure using two-stage boosting and a
higher compression ratio (CR).
l The use of internal EGR to increase the local charge
temperature in the combustion chamber when
needed via a variable valve timing (VVT) strategy.
High levels of EGR would be needed to control
engine-out NOx emissions so that both external
and internal EGR would be used with a trade-off in
local charge temperature between the competing
demands of lowering NOx emissions and achieving
stable combustion.
l The use of combustion assistance (e.g. a glow or
spark plug) to stabilise combustion at the lowest
load points.
The paper study also recognized the important role of
fuel spray and mixing, with higher pressure diesel injec-
tor systems being preferred along with an optimized
combustion chamber geometry.
Bench engine study
To test the learnings from the engineering paper study,
the bench engine study was carried out to provide a proof
of principle for the GCI engine concept, and to determine
what hardware measures, including ignition combustion
assistance, would be most effective for extending the
range of acceptable operation. The results from these
tests are presented here, based on the background pro-
vided in the ‘methodology’ section. A more detailed
account of the engine results is given in Rose et al.,2013.
The success criteria for the bench engine optimization
included the following factors: low engine-out NOx;
PM, HC and CO emissions as low as possible and suit-
able for further reduction by DOC (diesel oxidation cat-
alyst) and GPF (gasoline particulate filter) aftertreatment
systems; engine noise in the same range as conven-
tional diesel CI operation; and fuel efficiency at least as
good as the base engine configuration.
The bench engine included hardware enhancements
that enabled it to meet Euro VI emissions limits and
beyond. A downsizing concept was employed with a
Spark versus compression ignition in a new energy environment
13Volume 24, Number 1 • February 2015
cylinder pressureexhaust valve lift intake valve lift
180 360 540 720 900crank angle (degrees)
STAR-CD®
KIVA
mapping
Figure 2 Computational fluid dynamics (CFD) mapping methodology scheme—STAR-CD® to KIVA
cylinder swept volume of 390 cm3 that would allow the
construction of a 1.6-litre 4-cylinder engine while main-
taining the power of today’s 2.0-litre engines.
As anticipated by the engineering paper study, it proved
difficult to sustain reliable combustion using the market
gasoline at lower load operating conditions. Light-load
operation could be achieved, but NOx levels were
higher than desired. The combustion was also unstable
and would not tolerate additional EGR. For this reason,
the engine was fitted with a state-of-the-art glow plug
which was capable of a sustained glow temperature of
around 1200°C. For these tests, the engine coolant
temperature was also reduced to 48°C to simulate the
engine warm-up period.
The orientation of the glow plug to the injector spray is
known to be critical. The position was adjusted by
changing the orientation with respect to one individual
injector spray by one-degree increments, while moni-
toring engine performance. A position close to the
spray centre line giving the lowest CO/HC emissions
and combustion duration was chosen for testing.
With the glow plug installed, low-load operation was
possible at normal boost pressure levels, even at the
cooler engine temperature condition. Under hot engine
conditions, however, the glow plug did not help to
Spark versus compression ignition in a new energy environment
reduce the NOx emissions. At a 400-bar injection pres-
sure, combustion quality was poor with a higher EGR
rate. Reducing the injection pressure further to 260
bars improved combustion, but the increased heat
release led to higher NOx emissions even though the
EGR level was already quite high.
From the engineering paper study, it was expected that
the engine-out NOx/PM trade-off would be better com-
pared to diesel engines at low engine loads. Figure 3
shows the NOx levels achieved in this study for the var-
ious hardware options tested. The target NOx levels at
1500 rpm for various loads are shown by the solid
band marked ‘vehicle’. Even with the optimized injec-
tion strategy, higher CR, VVT and hot intake air, the
engine was not able to achieve the target NOx levels
without exceeding a reasonable level of HC emissions.
With combustion assist in the form of a glow plug, it
was possible to achieve loads down to 4.3 bars IMEP
as far as the combustion sound limit (CSL) would allow,
but not with the EGR levels required to meet the target
NOx levels.
In this study, internal (uncooled) EGR using negative
valve overlap was found to be advantageous for reduc-
ing HC emissions and improving fuel consumption in
the mid-load range.
There are a number of competing effects that occur
when more internal EGR is used. For example, higher
temperature by itself shortens the ignition delay but also
leads to higher NOx levels which require higher EGR
levels to control them. With higher EGR levels, the
decrease in local oxygen levels and inhomogeneities
associated with the internal EGR concentration led to
higher smoke levels and a tendency to lengthen the
ignition delay.
Nevertheless, Figure 4 (page 15) shows the brake-
specific fuel consumption (BSFC) results, for two differ-
ent speeds, obtained for the GCI concept compared
with the range of state-of-the-art 2012 model-year
spark ignition and direct injection compression ignition
engines. The GCI concept was found to be well below
those of naturally aspirated (NA) and turbo-charged
(TC) spark-ignited gasoline engines, and at the bottom
end of the range of direct injection diesel engines.
Concawe review14
IMEP (bar)
ISN
Ox
(g/k
Wh)
7
6
5
4
3
2
1
02 4 6 8 10 12 14 16
CR 19,HFR 310,hot intake air(critical CSL)
CR 17, HFR 310 and 520
CR 19, HFR 310 Vehicle
CR 19, HFR 310,hot intake air and adj. inj.; VVT
Figure 3 Indicated specific NOx (ISNOx) emissions achievable at 1,500 rpm as afunction of indicated mean effective pressure (IMEP)
In addition to optimizing the glow plug position and util-
ising internal EGR to improve gasoline’s ignitability
under the low load operating conditions, it was con-
cluded that further investigation would be required to
find the best configurations of VVT strategy and spray
targeting. Thus, three-dimensional CFD simulations
were carried out.
Modelling study
Because the engineering study suggested that a glow
plug would be required to assist combustion under this
low load operating condition, the spray spatial distribu-
tion and the local lambda (i.e. air/fuel ratio) were also
analysed.
Figure 5 shows the results of the mixture formation
analysis performed when the piston is at the top dead
centre firing (TDC-F) position for advanced boosted (left)
and more realistic boost conditions (right). Here, the
spray is visualized inside a 45° mesh sector of the piston
bowl, by means of rich lambda iso-volumes coloured by
Spark versus compression ignition in a new energy environment
15Volume 24, Number 1 • February 2015
engine speed (min-1)
BS
FC (g
/kW
h)
400
375
350
325
300
250
2001000 2000 3000 4000 5000
225
275
gasoline SI, TC, 35 engines
diesel DI, 33 engines
gasoline SI, NA, 31 engines
gasoline CI engine concept (this study)
Figure 4 Brake specific fuel consumption (BSFC) versus engine speed
Figure 5 Mixture formation under boosted (left) and more realistic (right) conditions
Reduced I/E-event cam shifting @ 48 °CA Reduced I/E-event cam shifting @ 48 °CA;diesel boundary conditions
700 800 900 1000 1100 1200
temperature (K)
increasing temperature. Thus, the rich zones, which will
probably ignite with the glow plug, are identified. With
accurate positioning of the glow plug, a favourable inter-
action between the fuel spray and glow plug is possible
with the chosen nozzle cone angle of 153°. With
advanced boosting the glow plug ignites spontaneously.
Under more normal boosting conditions the glow plug
will need to be heated to ignite the charge.
Further optimization was investigated using simulated
spark plug assist instead of a glow plug. Using the opti-
mum nozzle cone angle of 160° and nozzle protrusion
of >1.5 mm an ignitable mixture condition around the
spark plug is possible.
Apart from the work that Concawe has been doing over
a number of years, there is increasing interest in this
area of study from others. Delphi is working with
Hyundai in collaboration with the University of
Wisconsin-Madison’s Engine Research Consultants
(WERC) and Wayne State University on this topic, with
US Department of Energy funding. Argonne National
Laboratory and Saudi Aramco are also working on proj-
ects in this area of study. With a continuous and
increasing trend of gasoline oversupply, an engine that
performs at diesel engine fuel efficiency but running on
gasoline components seems to be a practical and
effective way forward to reduce costs and well-to-
wheels (WTW) emissions.
Acknowledgements
Concawe wishes to acknowledge FEV GmbH and
RWTH Aachen University who performed the work
described in this paper, and the members of its
Advanced Combustion Special Task Force (STF-26)
who contributed substantially to this study with the
support of Concawe’s member companies.
Spark versus compression ignition in a new energy environment
References
Cracknell, R., Ariztegui, J., Dubois, T., Hamje, H. et al.(2014). Modelling a Gasoline Compression Ignition (GCI)Engine Concept. SAE Technical Paper 2014-01-1305.
ExxonMobil (2014). The Outlook for Energy: A View to 2040.
Rose, K. D., Ariztegui, J., Cracknell, R., Dubois, T. et al.(2013). Exploring a Gasoline Compression Ignition (GCI)Engine Concept. SAE Technical Paper 2013-01-0911.
Concawe review16
17Volume 24, Number 1 • February 2015
Downstream oil industry safetystatistics for 2013
Safety management systems are widely recognized
by the oil industry as an essential tool for collecting
and analysing safety incident data and continuously
improving the safety of employees and contractors. To
support this effort, Concawe has, since 1993, been
compiling statistical safety data for the European down-
stream oil industry in order to:
1. provide member companies with a benchmark
against which to compare their own company’s
safety performance; and
2. demonstrate how responsible approaches to safety
management can help to ensure that accidents
stay at low levels in spite of the hazards that are
intrinsic to refinery and distribution operations.
Most importantly, Concawe’s annual safety data report
enables companies to evaluate the efficacy of their own
management systems, identify any shortcomings, and
take corrective actions as quickly as possible.
What safety data do we evaluate?
Concawe’s 20th report on the European downstream oil
industry’s safety performance (Concawe Report 8/14)
presents statistics on work-related personal injuries
sustained by oil industry employees and contractors
during 2013. It also highlights trends over the past 20
years of data collection and compares the oil industry’s
performance to that of other industrial sectors.
The 2013 report compiles safety data submitted by 34
Concawe member companies, representing about 93%
of the refining capacity of the EU-28 plus Norway and
Switzerland. The statistics are reported primarily in the
form of key performance indicators adopted by the
majority of oil companies operating in Europe, as well
as by other types of manufacturing industries. These
indicators are:
l Number of work-related fatalities;
l Fatal Accident Rate (FAR), expressed as the num-
ber of fatalities per 100 million hours worked;
l All Injury Frequency (AIF) expressed as the number
of injuries per million hours worked;
l Lost Workday Injuries (LWI) and the Lost Workday
Injury Frequency (LWIF) calculated by dividing the
LWI by the number of hours worked in millions;
l Lost Work Injury Severity (LWIS): the average num-
ber of lost workdays per LWI;
l Road Accident Rate (RAR): the number of road
accidents per million km travelled; and
l Process Safety Performance Indicators (PSPIs) that
report the number of Process Safety Events (PSEs)
expressed as unintended Losses of Primary
Containment (LOPCs).
Process Safety Performance Indicators
Several major industrial incidents, including the
Toulouse explosion (2001), the Buncefield fire (2005)
and the Texas refinery explosion (2005), have led to
increased attention being given to the causation of
such events. This has led to several initiatives that
focus on the gathering of PSPIs. The lagging indicator
for this is the PSEs, mainly Losses of Primary
Containment because these have been proven to be
the initiating events for the aforementioned disasters.
PSPI data were collected in 2013 for the fifth consecu-
tive year, following the publication of the latest recom-
mended practice of the American Petroleum Institute
(API). The additional data provide insights into the types
and causes of process safety incidents. PSPIs also
enable the refining and distribution industry to compare
their European process safety performance with similar
data from other regions of the world.
Thirty-two Concawe companies provided PSPI data in
2013. From these responses, a Process Safety Event
Rate (PSER) indicator of 1.7 was recorded for all PSEs,
which is the lowest result ever. The overall results of
the PSPI survey are presented in Table 1 (overleaf).
The 2013 safety
statistics report
presents data on
personal injuries and
process safety,
highlighting trends
over the past 20 years
of data collection.
©S
hutt
erst
ock.
com
Downstream oil industry safety statistics for 2013
Concawe review18
Fortunately, none of the reported PSEs resulted in a
major incident that the understanding of PSE causation
is trying to prevent.
Since the PSI data gathering was started in 2009, there
has been a gradual decrease in the PSER, irrespective of
the number of reporting companies, as can be seen in
Figure 1. This decreasing trend is a good example of the
commitment of the Concawe membership to process
safety management, and furthermore demonstrates that
the systematic gathering of such data enables the mem-
bership to actively manage this operational threat.
Personal Safety Indicators
Accident frequencies in the European downstream oil
industry have been historically quite low; the 2013 data
shows a 1.1 LWIF for 2013, which is the lowest value
ever reported in the sector.
In general, performance indicator results are of greatest
interest when these can be analysed for historical
trends. The evolution of safety performance over a
period of time provides indications on how well safety
management efforts are working. Figure 2, for example,
shows the changes and improving trends in the three-
year rolling averages for the four main performance
indicators mentioned above.
The trends in these indicators show a steady performance
improvement over the past 20 years, with a slow but con-
stant reduction in LWIF which remained below 2.0 for the
fifth consecutive year. Although the data suggest that AIF
peaked around 1996–97, this could also result from better
data reporting as the AIF indicator was not formally used
in all companies in the early years of Concawe’s data
gathering. Since 1997, the trend in AIF has generally been
downwards except for a slight increase in 2010.
Regrettably, six fatalities in five separate incidents were
reported in 2013:
l one of these fatalities was due to a road accident;
l three were due to two pressure release incidents;
l one was caused by a worker caught in, under or
between a moving mass; and
l one was caused by a fall.
Table 1 Results of the 2013 PSPI survey
Sector Manufacturing Marketing Both sectors
Companies reportingTotal 39 23 22Process safety reporting 32 13 13Percentage 82% 57% 59%
Hours worked (Mh)Total 281 292.5 573.5Process safety reporting 268.2 a 223.1 457.7Percentage 95%a 76% 86%
Tier 1 PSE: PSE 115 9 124
Tier 2 PSE: PSE 334 81 415
Tier 1 PSER: PSE/Mh reported 0.43 0.04 0.27
Tier 2 PSER: PSE/Mh reported 1.25 0.36 1.34
Total PSER: PSE/Mh reported 1.67 0.40 1.18
a All companies provided both Tier 1 and Tier 2 PSEs for 2013.
Figure 1 PSE data for manufacturing, 2009–2013
PS
ER
4.5
3.5
3.0
2.0
1.0
0
0.5
2009 2011
PSER-1 PSER-2
2010
2.5
1.5
35
30
20
10
0
5
25
15
No.
of C
onca
we
mem
bers
repo
rtin
g
4.0
PSER number of Concawe members reporting
2012 2013reporting year
1.09
2.99
4.08
0.87
2.71
3.57
0.71
3.20
3.81
0.59
2.20
2.63
0.49
1.42
1.91
num
ber
of in
cide
nts
(3-y
ear
rollin
g av
erag
es)
0
2
4
6
10
12
8
1993
–95
1994
–96
1995
–97
1996
–98
1998
–00
1999
–01
2000
–02
1997
–99
2001
–03
2002
–04
2003
–05
2004
–06
2005
–07
2006
–08
2008
–10
2007
–09
2009
–11
2010
–12
2011
–13
RAR: Road Accident Rate (road accidents per million km travelled)
AIF: All Injury Frequency (injuries per million hours worked)
FAR: Fatal Accident Rate (fatalities per 100 million hours worked)
LWIF: Lost Workday Injury Frequency (lost time injuries per million hours worked)
Figure 2 Three-year rolling average personal incident statistics for the Europeandownstream oil industry
Downstream oil industry safety statistics for 2013
19Volume 24, Number 1 • February 2015
The six fatalities in 2013 are again the lowest ever expe-
rienced since Concawe started to collect safety data
(Figure 3). After a steady downward trend during the
1990s, fatalities began to increase again in 2000 with a
very high value of 22 fatalities in 2003. This
unfavourable trend was reversed in 2004–06 and the
fatality numbers have shown little variation since that
time. The three-year rolling average for FAR has also
stayed at about 2 for the past four years.
In 2013, contractors in the manufacturing sector of the
European oil industry were the most vulnerable work
group, experiencing four fatalities. This clearly remains
a concern and demonstrates that all companies should
ensure that their contractor workforce is fully integrated
into their safety awareness and monitoring systems.
The relationships between the AIF, LWIF and FAR are
presented in Figure 4.
While the number of fatalities per year has an impact on
the two curves that are associated with FAR values, the
figure shows relatively stable relationships among these
indicators over time. Almost half of safety incidents are
LWIs and there was approximately one fatality for every
100 LWIs.
Contrary to the positive trends in the LWIF and AIF indi-
cators, the LWIS indicator, expressing the average
number of days lost per LWI, increased in 2013. LWIS
data and the three-year rolling average are shown in
Figure 5. Although the LWIS results declined after peak-
ing in 2010, the three-year rolling average still remains
above the all-time LWIS average of 25. Therefore, the
severity of the incidents that occur remains a concern.
Causes of fatalities and LWIs
In the 2013 survey, Concawe also gathered information
on the causes of Lost Work Injuries in order to see how
closely the LWIs could be related to the causes of fatal-
ities. In 2013 the LWIs were categorised in five main
categories also used to report the causes of the fatali-
ties. These five categories were selected after ample
analysis of the reporting method for this kind of data by
other industrial sectors, and of the previous practice
within the Concawe membership. The result is a scheme
Figure 3 Numbers of reported fatalities since 1993
num
ber
of fa
talit
ies
0
5
10
15
25
20
1993
1994
1995
1996
1998
1999
2000
1997
2001
2002
2003
2004
2005
2006
2008
2007
2009
2010
2011
2012
2013
reporting year
owntotal contract
4
3
2
1
0
5
1992 1994 1996 1998 2000 2002 2004 2006 2008 2014
6
num
ber
of in
cide
nts
2010 2012
LWIF/FARAIF/LWIFAIF/FAR
days
lost
per
LW
I
20
25
45
30
1993
1995
1996
1997
1999
2000
2001
1998
2002
2003
2004
2005
2006
2007
2008
2013
15
10
5
1994
2009
35
40
2010
2011
2012
0
LWIS (average number of days lost per LWI) 3-year rolling average
Figure 4 Relationships between incidents and fatalities for the Europeandownstream oil industry
Figure 5 Lost Workday Injury Severity (LWIS) from 1993–2013 and the three-yearrolling average for the European downstream oil industry
Downstream oil industry safety statistics for 2013
Concawe review20
Table 2 LWIs and their causes
Cause Manufacturing Marketing Combined Percentage
Road accident Road accidents 6 22 28 4%
Height/falls Falls from height 24 42 66 10%
Staff hit by falling objects 10 13 23 4%
Slips and trips (same height) 89 121 210 33%
Burn/electrical Explosion or burns 28 3 31 5%
Exposure (electrical) 1 3 4 1%
Confined space Confined space 3 2 5 1%
Other causes Assault or violent act 0 11 11 2%
Water-related, drowning 1 0 1 0%
Cut, puncture, scrape 11 21 32 5%
Struck by 37 25 62 10%
Exposure, noise, chemical, 16 1 17 3% biological, vibration
Caught in, under or between 28 19 47 7%
Overexertion, strain 31 49 80 12%
Pressure release 6 0 6 1%
Other 9 11 20 3%
Total 300 343 643 100%
very closely related to that of the International
Association of Oil & Gas Producers (IOGP), an associ-
ation comprising many Concawe members and per-
forming scientific advocacy on behalf of their
exploration and production activities.
A total of 643 LWIs were reported in 2013, of which
only 20 (3%) could not be assigned to one of the 5
agreed categories by the reporting member compa-
nies. An overview of the LWI incidents and causes are
provided in Table 2.
As can be seen from Figure 6, the percentage data for
these LWIs in 2013 show that the distribution of LWI
causes is quite different from those that resulted in
fatalities.
This data being relatively new, there is no basis yet for
a robust analysis of trends. Concawe will continue to
collect this data in future years and the results should
reveal trends that can be analysed in greater depth,
providing valuable data to member companies that can
then be used to improve on-the-job safety for employ-
ees and contractors.
Figure 6 Reported causes on a percentage basis for LWIs and fatalities in 2013
burns/electrical
confined space
pressure release
road incidents
height/falls
other
60%
40%
20%
0%
LWIsfatalities
Abbreviations and terms
21Volume 24, Number 1 • February 2015
AIF All Injury Frequency
API American Petroleum Institute
BAT Best Available Techniques
BAT REF BAT Reference document. Full title:or BREF ‘Reference Document on Best Available
Techniques for ….’ (A series of documentsproduced by the European IntegrationPollution Prevention and Control Bureau(EIPPCB) to assist in the selection of BATsfor each activity area listed in Annex 1 ofDirective 96/61/EC)
BSFC Brake-Specific Fuel Consumption
BSI British Standards Institution
CA Crank Angle
CFD Computational Fluid Dynamics
CI Compression Ignition
CNG Compressed Natural Gas
CO Carbon Monoxide
CR Compression Ratio
CSL Combustion Sound Limit
DI Direct Injection
DOC Diesel Oxidation Catalyst
EGR Exhaust Gas Recirculation
FAR Fatal Accident Rate
FID Flame Ionisation Detector
GCI Gasoline Compression Ignition
GHG Greenhouse Gas
GPF Gasoline Particulate Filter
GT-Power Part of the ‘GT-Suite’ of engineeringsoftware, and the industry standard forengine simulations
HC Hydrocarbon
HFS High Flow® Sampler—a device designed forestimating leaks of natural gas
I/E Intake/Exhaust
IMEP Indicated Mean Effective Pressure
IOGP International Association of Oil & Gas Producers
ISNOx Indicated Specific NOx
KIVA A family of Fortran-based CFD softwaredeveloped by the Los Alamos NationalLaboratory
LDAR Leak Detection and Repair
LNG Liquefied Natural Gas
LOPC Loss Of Primary Containment
LTC Low-Temperature Combustion
LWI Lost Workday Injury
LWIF Lost Workday Injury Frequency
LWIS Lost Workday Injury Severity
Method 21 A methodology for identifying leakingequipment by using a hydrocarbonionisation detector
NA Naturally Aspirated
NEN Netherlands Standardization Institute
NMVOC Non-Methane Volatile Organic Compounds
NOx Nitrogen Oxides
OGI Optical Gas Imaging
PID Photo Ionisation Detector
PPMV Parts Per Million by Volume
PSE Process Safety Event
PSER Process Safety Event Rate
PSPI Process Safety Performance Indicator
RAR Road Accident Rate
SI Spark Ignition
TDC-F Top Dead Centre Firing (position)
TC Turbo-Charged
US-EPA United States Environmental ProtectionAgency
VOC Volatile Organic Compound
VVT Variable Valve Timing
WTW Well To Wheels
Concawe contacts
Concawe review22
Chris BeddoesTel: +32-2 566 91 05E-mail: [email protected]
Robin NelsonTel: +32-2 566 91 61 Mobile: +32-496 27 37 23E-mail: [email protected]
Air qualityLucia Gonzalez BajosTel: +32-2 566 91 71 Mobile: +32-490 11 04 71E-mail: [email protected]
Fuels quality and emissionsHeather HamjeTel: +32-2 566 91 69 Mobile: +32-499 97 53 25E-mail: [email protected]
HealthArlean RohdeTel: +32-2 566 91 63 Mobile: +32-495 26 14 35E-mail: [email protected]
Petroleum products • Risk assessmentFrancisco del Castillo RomanTel: +32-2 566 91 66 Mobile: +32-490 56 84 83 E-mail: [email protected]
Petroleum products • SafetyKlaas den HaanTel: +32-2 566 91 83 Mobile: +32-498 19 97 48E-mail: [email protected]
REACH Implementation Manager & Legal AdvisorSophie BornsteinTel: +32-2 566 91 68 Mobile: +32-497 26 08 05E-mail: [email protected]
Refinery technology Alan ReidTel: +32-2 566 91 67 Mobile: +32-492 72 91 76E-mail: [email protected]
Water, soil and waste • Oil pipelinesMike SpenceTel: +32-2 566 91 80 Mobile: +32-496 16 96 76E-mail: [email protected]
Office Support
Marleen Eggerickx Tel: +32-2 566 91 76E-mail: [email protected]
Sandrine FaucqTel: +32-2 566 91 75E-mail: [email protected]
Jeannette HenriksenTel: +32-2 566 91 05E-mail: [email protected]
Anja MannaertsTel: +32-2 566 91 73E-mail: [email protected]
REACH Support
Jessica Candelario PerezTel: +32-2 566 91 65E-mail: [email protected]
Julie TorneroTel: +32-2 566 91 73E-mail: [email protected]
Finance, Administration & HR ManagerDidier De VidtsTel: +32-2 566 91 18 Mobile: +32-474 06 84 66E-mail: [email protected]
Finance, Administration & HR Support
Alain LouckxTel: +32-2 566 91 14E-mail: [email protected]
Madeleine DasnoyTel: +32-2 566 91 37E-mail: [email protected]
Communications ManagerAlain MathurenTel: +32-2 566 91 19E-mail: [email protected]
Communications Support
Lukasz PasterskiTel: +32-2 566 91 04E-mail: [email protected]
Director General Office management and support
Science Director
Science Executives
Air qualityKaisa Vaskinen
Refining and fuelsCatarina Caiado
Research Associates
Concawe publications
23Volume 24, Number 1 • February 2015
Reports published by Concawe from 2014 to date
Adobe PDF files of virtually all current reports, as well as up-to-date catalogues, can be downloaded from Concawe’s website at:
www.concawe.org/content/default.asp?PageID=569.
Readers can receive notification when a new report is published by subscribing to the relevant RSS feeds:
www.concawe.org/content/default.asp?PageID=636.
12/14 Towards the establishment of a protocol for the quantification of VOC diffuse emissions using open-path remotemonitoring techniques: DIAL monitoring of a VOC source of known emission flux
11/14 The estimated forward cost of EU legislation for the EU refining industry
10/14 Hazard classification and labelling of petroleum substances in the European Economic Area – 2014
9/14 Review of recent health effect studies with nitrogen dioxide
8/14 European downstream oil industry safety performance – statistical summary of reported incidents – 2013
7/14 Impact of FAME on the performance of three Euro 4 light-duty diesel vehicles – Part 2: Unregulated emissions
6/14 Impact of FAME on the performance of three Euro 4 light-duty diesel vehicles – Part 1: Fuel consumption and regulatedemissions
5/14 Methods for estimating VOC emissions from primary oil-water separator systems in refineries
4/14 Use of motor fuels and lubricants: habits and practices of consumers in Europe
3/14 Assessment of Recent Health Studies of Long-Term Exposure to Ozone
2/14 Proceedings of the Mineral Oil CRoss INdustry IssueS (MOCRINIS) Workshop – September 2013
1/14 Application of the target lipid and equilibrium partitioning models to non-polar organic chemicals in soils and sediments
2014 JEC Biofuels Study: 2014 update. EU renewable energy targets in 2020: Revised analysis of scenarios for transportfuels report
2014 JRC/EUCAR/CONCAWE (JEC) Well-to-Wheels studies. The JEC Consortium periodically updates its Well-to-Wheelsstudies that evaluate greenhouse gas emissions and energy efficiency of various energy pathways, automotive fuels, andvehicle powertrain options that are relevant to Europe from 2010 to 2020+. Version 4 Well-to-Wheels (WTW) Report(March 2014) can be downloaded from the JRC website.
Other reports co-authored by Concawe from 2014 to date
This page is intentionally blank
This page is intentionally blank
Concawe
Boulevard du Souverain 165, B–1160 Brussels, Belgium
Telephone: +32-2 566 91 60 • Telefax: +32-2 566 91 81
[email protected] • www.concawe.org